Comments --
1. Of course the Bogen design already includes control grid stopper resistors, so the article must be referring to the lack of -- and therefore the installation of -- screen stopper resistors. Except that due to the physical design of the 8417 tube (cavity anode), use of screen stopper resistors has little if any impact at all on parasitic oscillations with this tube. This was (again) proven out back when I was involved in the project of this thread, where you will note that for the 2 X 8417 version, the plate stopper resistors were absolutely required, and left in as a result, but of course the 2 X 6550 version required no such resistors. And, because the screens were operating at a much lower voltage than the plate, and because there were only two tubes involved, the 2 X 6550 version actually requires no stopper resistors at all other than the garden variety control grid stoppers. This is one reason that the 6550 version is capable of more power output than the 8417 version.
2. Understand that in this type of amplifier design, it is impossible for the output tube control grids to ever be driven positive. They can be driven right up to Eg1=0, but never above that level. That's because at any potential above Eg1=0, the control grid starts to conduct current, just as the screen grid of plate elements do by operating at a positive level relative to the cathode. To make the control grids go positive requires a driver stage than can actually deliver power (a combination of voltage and current), which then results in Class B2 operation. However, the little driver of these amplifiers is nowhere near capable of delivering the necessary drive power (which can be of a level of a few watts in some cases) to produce Class B2 operation, so the design firmly remains a Class B1 design, with the control grids never being driven positive. As well, the 1000Ω control grid stopper resistors would preclude Class B2 operation in any of these related designs anyway. Designs that operate or ultimate shift in/to Class B2 always have very low DC resistance values in the grid #1 circuit, precisely so that enough current can flow to actually force the control grid to a positive level. Back in the day, they always used a driver transformer to accomplish this task while today, mosfets can do the job.
3. These industrial designs invariably will always use a separate or tertiary winding to supply the NFB loop. The purpose of this is so that the load winding can operate independent of a ground connection. Were the load winding to supply the NFB signal, then the winding must be grounded at some point to complete this circuit. But when the signal is supply by a separated -- grounded -- winding, then the main winding can be grounded or not, as local codes or the installation requires. Remember (again), these amplifiers were designed for a purpose, which could be anything from a large sound distribution network, to powering shaker tables. Some installations required the load to operate independent of ground, while others required it to operate with one side grounded. The use of a separate FB winding facilitated this ability. In the better OPTs, (which the Bogens are), this tertiary winding will be tightly wound with the load winding, so that the FB winding produces a faithful -- but isolated -- copy of signal appearing at the load winding.
4. The NFB winding in the Bogen is a 70 volt winding as well. The voltage level appearing there at maximum output merely represents a design exercise so that the proper amount of FB voltage is applied back to the inverting input of the amplifier.
5. I'm not sure if I ever directly measured the distortion of the original design OL -- but I may have in chasing down the phase inverter resistor issue. Check back around pages 8-10.
6. Fender and others aside, it is never a good idea to leave vacuum tubes with their heaters powered, but the tubes passing no current. This can lead to damaging the cathode which is fully documented, primarily with the early studies involving vacuum tube based computers. Providing a standby feature is always a good idea when it can be used to advantage, but when activated, it should still allow for a small amount of current flow (even 5 mA is sufficient) to prevent cathode damage.
7. Adding a cathode bypass cap to the AF Amplifier stage (requiring a slight re-design as to how the NFB is applied) will add very little if any additional gain at all to this design. That's because the cathode resistor is already so small in this design -- relative to the large plate resistor value employed -- that there is virtually no current feedback developed across the resistor for a bypass cap to bypass. In other words, they've already extracted as much gain out of the stage as possible, as further evidenced by the very low amount of static bias voltage -- and the problem that causes -- that I've previously documented. On the other hand, you will note that in the modified 2 X 8417 version I developed, the cathode resistor of the AF Amplifier stage is raised significantly -- this to solve the non-linear operation of the phase inverter at high power levels -- but that also then introduces current feedback into the stage. As a result, I bypassed that resistor to achieve both a more properly level of static bias, and the AC gain of the original design.
8. Bogen's power supply approach is a more economical way to approach the needs of the amplifier, and in practice works well enough. The Stromberg design however has the advantage of operating the screen grids from a full-wave rectifier design, which minimizes any filtering required. The Bogen approach operates the screens from a half-wave supply, requiring the need for filtering, which then introduces power robbing impedance into the design. This approach is a significant factor in 8417 designs, as the screen grid of this tube draws a significant amount of current -- much more than the aligned grid 6550 beam power tube does, allowing more current to flow to the plate, and is yet another reason it is capable of producing more power output.
Dave
1. Of course the Bogen design already includes control grid stopper resistors, so the article must be referring to the lack of -- and therefore the installation of -- screen stopper resistors. Except that due to the physical design of the 8417 tube (cavity anode), use of screen stopper resistors has little if any impact at all on parasitic oscillations with this tube. This was (again) proven out back when I was involved in the project of this thread, where you will note that for the 2 X 8417 version, the plate stopper resistors were absolutely required, and left in as a result, but of course the 2 X 6550 version required no such resistors. And, because the screens were operating at a much lower voltage than the plate, and because there were only two tubes involved, the 2 X 6550 version actually requires no stopper resistors at all other than the garden variety control grid stoppers. This is one reason that the 6550 version is capable of more power output than the 8417 version.
2. Understand that in this type of amplifier design, it is impossible for the output tube control grids to ever be driven positive. They can be driven right up to Eg1=0, but never above that level. That's because at any potential above Eg1=0, the control grid starts to conduct current, just as the screen grid of plate elements do by operating at a positive level relative to the cathode. To make the control grids go positive requires a driver stage than can actually deliver power (a combination of voltage and current), which then results in Class B2 operation. However, the little driver of these amplifiers is nowhere near capable of delivering the necessary drive power (which can be of a level of a few watts in some cases) to produce Class B2 operation, so the design firmly remains a Class B1 design, with the control grids never being driven positive. As well, the 1000Ω control grid stopper resistors would preclude Class B2 operation in any of these related designs anyway. Designs that operate or ultimate shift in/to Class B2 always have very low DC resistance values in the grid #1 circuit, precisely so that enough current can flow to actually force the control grid to a positive level. Back in the day, they always used a driver transformer to accomplish this task while today, mosfets can do the job.
3. These industrial designs invariably will always use a separate or tertiary winding to supply the NFB loop. The purpose of this is so that the load winding can operate independent of a ground connection. Were the load winding to supply the NFB signal, then the winding must be grounded at some point to complete this circuit. But when the signal is supply by a separated -- grounded -- winding, then the main winding can be grounded or not, as local codes or the installation requires. Remember (again), these amplifiers were designed for a purpose, which could be anything from a large sound distribution network, to powering shaker tables. Some installations required the load to operate independent of ground, while others required it to operate with one side grounded. The use of a separate FB winding facilitated this ability. In the better OPTs, (which the Bogens are), this tertiary winding will be tightly wound with the load winding, so that the FB winding produces a faithful -- but isolated -- copy of signal appearing at the load winding.
4. The NFB winding in the Bogen is a 70 volt winding as well. The voltage level appearing there at maximum output merely represents a design exercise so that the proper amount of FB voltage is applied back to the inverting input of the amplifier.
5. I'm not sure if I ever directly measured the distortion of the original design OL -- but I may have in chasing down the phase inverter resistor issue. Check back around pages 8-10.
6. Fender and others aside, it is never a good idea to leave vacuum tubes with their heaters powered, but the tubes passing no current. This can lead to damaging the cathode which is fully documented, primarily with the early studies involving vacuum tube based computers. Providing a standby feature is always a good idea when it can be used to advantage, but when activated, it should still allow for a small amount of current flow (even 5 mA is sufficient) to prevent cathode damage.
7. Adding a cathode bypass cap to the AF Amplifier stage (requiring a slight re-design as to how the NFB is applied) will add very little if any additional gain at all to this design. That's because the cathode resistor is already so small in this design -- relative to the large plate resistor value employed -- that there is virtually no current feedback developed across the resistor for a bypass cap to bypass. In other words, they've already extracted as much gain out of the stage as possible, as further evidenced by the very low amount of static bias voltage -- and the problem that causes -- that I've previously documented. On the other hand, you will note that in the modified 2 X 8417 version I developed, the cathode resistor of the AF Amplifier stage is raised significantly -- this to solve the non-linear operation of the phase inverter at high power levels -- but that also then introduces current feedback into the stage. As a result, I bypassed that resistor to achieve both a more properly level of static bias, and the AC gain of the original design.
8. Bogen's power supply approach is a more economical way to approach the needs of the amplifier, and in practice works well enough. The Stromberg design however has the advantage of operating the screen grids from a full-wave rectifier design, which minimizes any filtering required. The Bogen approach operates the screens from a half-wave supply, requiring the need for filtering, which then introduces power robbing impedance into the design. This approach is a significant factor in 8417 designs, as the screen grid of this tube draws a significant amount of current -- much more than the aligned grid 6550 beam power tube does, allowing more current to flow to the plate, and is yet another reason it is capable of producing more power output.
Dave
